This document describes the various replication files associated with:
Jacoby, William G. (2014) "Is There a Culture War? Conflicting Value
Structures in American Public Opinion." AMERICAN POLITICAL SCIENCE REVIEW
108(4): 754-771. 
DOI: http://dx.doi.org.proxy1.cl.msu.edu/10.1017/S0003055414000380


  I. ORIGINAL 2006 CCES DATASET USED FOR THE ANALYSIS

     "MSU_VARLIST.pdf"

        Codebook for file containing Michigan State 
        University module and Common Content from
        the 2006 Cooperative Congressional Election
        Study. This dataset is the original source for
        all analyses reported in the APSR article.


     "MSU, Common, CCES Dataset.sav"

        SPSS dataset containing Michigan State
        University module and Common Content from 
        the 2006 Cooperative Congressional Election
        Study.


     "msu_common_cces_dataset.sas7bdat"

        SAS dataset containing Michigan State
        University module and Common Content from 
        the 2006 Cooperative Congressional Election
        Study.



 II. FILES USED TO ESTIMATE THE VECTOR MODEL

     "Estim Nonmetric Vector Model.R"

        This file contains the R session that: (1) reads in 
        the data from the 2006 Cooperative Congressional 
        Election Study (CCES) on respondents' rank-ordered 
        importance ratings for seven values; (2) estimates 
        the nonmetric MDPREF model; (3) plots the estimated
        model; (4) calculates the mean vector; and (5) writes 
        out a text file containing the CCES case ID and
        the vector coordinates for each respondent.


     "ranked values, with names.txt"

        Text file containing one line for each of the 775 
        respondents from the 2006 CCES who provided complete
        rank-orders for the seven values. The file begins
        with a header record that provides variable names.
        On each line, adjacent data values are separated by
        whitespace. The variable names and definitions are 
        as follows (in the order that they appear on
        each line of the data file):

          caseid  Case identification number from the 
                  2006 CCES

          rankfr  Number of values the respondent ranks as
                  less important than freedom

          rankeq  Number of values the respondent ranks as
                  less important than equality

          rankes  Number of values the respondent ranks as
                  less important than economic security

          rankmo  Number of values the respondent ranks as
                  less important than morality

          rankin  Number of values the respondent ranks as
                  less important than individualism

          rankso  Number of values the respondent ranks as
                  less important than social order

          rankpa  Number of values the respondent ranks as
                  less important than patriotism


     "optiscale_1.0-4.tar.gz"

        Archive containing the "optiscale" R package. This
        package is required in order to estimate the 
        nonmetric vector model. It is used in the R session
        that is contained in the file, "Estim Nonmetric
        Vector Model.R". The "optiscale" package
        is also available on CRAN.


     "row vectors.txt"

        Text file containing the coordinates of the terminal
        points of the vectors representing the respondents
        to the 2006 CCES in the nonmetric MDPREF model. This
        file is written from the R session that is contained
        in the file, "Estim Nonmetric Vector Model.R". There
        is one line for each of the 775 CCES respondents and
        on each line, adjacent data values are separated by
        whitespace. The variables are as follows (in the order
        that they appear on each line of the data file):

          Case identification number from the 2006 CCES

          First dimension coordinate for terminal point
          of the respondent vector

          Second dimension coordinate for terminal point
          of the respondent vector


     "value coords, with names.txt"

        Text file containing the coordinates of the 
        points representing the values, obtained from
        the nonmetric MDPREF model estimated in the R
        session contained in the file, "Estim Nonmetric 
        Vector Model.R". There is one line for each value
        and on each line, adjacent data values are separated
        by whitespace. The first line of the file is a header
        record containing variable names. The variables are 
        as follows (in the order that they appear on each 
        line of the data file). 

          Value name

          First dimension coordinate for the value point

          Second dimension coordinate for the value point


III. FILES USED TO CREATE FIGURE 3

     "relig.sas7bdat"

        SAS dataset containing variables and observations
        extracted from "msu_common_cces_dataset.sas7bdat"
        in order to construct Figure 3. The relevant 
        variables in this dataset are:

          caseid  Case identification number from CCES 
                  dataset (CCES variable name is CaseID)

        religion  Religious affiliation variable from
                  cces dataset (CCES variable name
                  is religion)

            dim1  First dimension coordinates for 
                  terminal points of individual 
                  respondent vectors, Merged with CCES
                  data from "row vectors.txt"

            dim2  Second dimension coordinates for 
                  terminal points of individual 
                  respondent vectors, Merged with CCES
                  data from "row vectors.txt"

          commit  Summary measure of religious
                  commitment, created from CCES
                  variables pr_relig1, pr_relig3,
                  pr_relig4, pr_relig5, and
                  pr_relig6. These variables were 
                  recoded with successive-integer
                  scores, starting with zero. In each
                  variable, larger integers correspond
                  to greater religious commitment. The
                  variables were used in a nonparametric
                  IRT model (Mokken scale), where it
                  was determined that CCES items pr_relig2
                  and pre_relig7 did not scale with the
                  other variables. The commitment variable
                  scores are obtained by summing the recoded
                  values for pr_relig1, pr_relig3-pr_relig6.


     "relig.dta"

        Stata dataset containing the same information as 
        "relig.sas7bdat" 

     
     "Create Figure 3.R"

        File containing R session that uses data from 
        "relig.dta" to create Figures 3A and 3B, and 
        perform analysis of angular variation (ANAVA)
        on religious denominations and the four-category
        version of the religious commitment variable.



 IV. FILES USED TO CREATE FIGURE 4

     "info.sas7bdat"

        SAS dataset containing variables and observations
        extracted from "msu_common_cces_dataset.sas7bdat"
        in order to construct Figure 4. The relevant 
        variables in this dataset are:

          caseid  Case identification number from CCES 
                  dataset (CCES variable name is CaseID)

            dim1  First dimension coordinates for 
                  terminal points of individual 
                  respondent vectors, Merged with CCES
                  data from "row vectors.txt"

            dim2  Second dimension coordinates for 
                  terminal points of individual 
                  respondent vectors, Merged with CCES
                  data from "row vectors.txt"

        ideolpst  Ideological self-placement on 
                  0-100 scale, from post-election
                  wave of the 2006 CCES (CCES
                  variable name is postQ10)

        postpid7  Seven-point party identification
                  scale, constructed from the standard
                  branching questions included in
                  the post-election wave of the
                  2006 CCES (CCES variables 
                  postQ15, postQ10a, and postQ10b)


     "info.dta"

        Stata dataset containing the same information
        as "info.sas7bdat"


     "Create Figure 4.R"

        File containing R session that uses data from 
        "info.dta" to create Figures 4A and 4B, and 
        perform analysis of angular variation (ANAVA)
        on three-category versions of party identification
        and ideological self-placement variables.



  V. FILES USED TO CREATE TABLE 3 AND FIGURE 5

     "circreg.sas7bdat"

        SAS dataset containing variables and observations
        extracted from "msu_common_cces_dataset.sas7bdat"
        in order to perform the circular regression reported
        in Table 3 and to construct Figure 5. The relevant 
        variables in this dataset are:

          caseid  Case identification number from CCES 
                  dataset (CCES variable name is CaseID)

            dim1  First dimension coordinates for 
                  terminal points of individual 
                  respondent vectors, Merged with CCES
                  data from "row vectors.txt"

            dim2  Second dimension coordinates for 
                  terminal points of individual 
                  respondent vectors, Merged with CCES
                  data from "row vectors.txt"

        ideolpst  Ideological self-placement on 
                  0-100 scale, from post-election
                  wave of the 2006 CCES (CCES
                  variable name is postQ10)

        postpid7  Seven-point party identification
                  scale, constructed from the standard
                  branching questions included in
                  the post-election wave of the
                  2006 CCES (CCES variables 
                  postQ15, postQ10a, and postQ10b)

          commit  Summary measure of religious
                  commitment, created from CCES
                  variables pr_relig1, pr_relig3,
                  pr_relig4, pr_relig5, and
                  pr_relig6. These variables were 
                  recoded with successive-integer
                  scores, starting with zero. In each
                  variable, larger integers correspond
                  to greater religious commitment. The
                  variables were used in a nonparametric
                  IRT model (Mokken scale), where it
                  was determined that CCES items pr_relig2
                  and pre_relig7 did not scale with the
                  other variables. The commitment variable
                  scores are obtained by summing the recoded
                  values for pr_relig1, pr_relig3-pr_relig6.

             age  Respondent's age. Created from CCES
                  variable birthyr by taking
                  2006 - birthyr

     somecollege  Dummy variable coded 1 for respondents who
                  reported some college or a two-year degree.
                  and zero otherwise. Created from CCES 
                  variable educ, categories 3 and 4.

     collegegrad  Dummy variable coded 1 for respondents who
                  graduated from college, and zero otherwise.
                  Created from CCES variable educ, 
                  categories 5 and 6
  

     "circreg.dta"

        Stata dataset containing the same information
        as "circreg.sas7bdat"


     "Run Circ Reg and Create Figure 5.R"

        File containing R session that uses data from
        "circreg.dta" to run the circular regression 
        that is reported in Table 3, and to create
        Figure 5